The Determinants of Exit in Argentina:

Core and Peripheral Regions

Carla Daniela Calá Miguel Manjón-Antolín 

Josep Maria Arauzo-Carod

Document de treball n.04- 2014

WORKING PAPERS

Col·lecció “DOCUMENTS DE TREBALL DEL DEPARTAMENT D’ECONOMIA - CREIP”

DEPARTAMENT D’ECONOMIA – CREIP Facultat d’Economia i Empresa

UNIVERSITAT

VROVIRA I IRGILI

DEPARTAMENT D’ECONOMIA

Edita:

Adreçar comentaris al Departament d’Economia / CREIP Dipòsit Legal: T - 461 - 2014 ISSN edició en paper: 1576 - 3382 ISSN edició electrònica: 1988 - 0820

DEPARTAMENT D’ECONOMIA – CREIP

Facultat d’Economia i Empresa

Departament d’Economia www.fcee.urv.es/departaments/economia/public_html/index.html Universitat Rovira i Virgili Facultat d’Economia i Empresa Avgda. de la Universitat, 1 43204 Reus Tel.: +34 977 759 811 Fax: +34 977 300 661 Email: sde@urv.cat

CREIP www.urv.cat/creip Universitat Rovira i Virgili Departament d’Economia Avgda. de la Universitat, 1 43204 Reus Tel.: +34 977 558 936 Email: creip@urv.cat

UNIVERSITAT

VROVIRA I IRGILI

DEPARTAMENT D’ECONOMIA

The Determinants of Exit in Argentina: Core and Peripheral Regions

Carla Daniela Calá (): dacala@mdp.edu.ar Miguel Manjón-Antolín (): miguel.manjon@urv.cat

Josep-Maria Arauzo-Carod (,♦): josepmaria.arauzo@urv.cat

Abstract 

This paper analyses the regional determinants of exit in Argentina. We find evidence of a dynamic revolving door by which past entrants increase current exits, particularly in the peripheral regions. In the central regions, current and past incumbents cause an analogous displacement effect. Also, exit shows a U-shaped relationship with respect to the informal economy, although the positive effect is weaker in the central regions. These findings point to the existence of a core-periphery structure in the spatial distribution of exits. Key words: firm exit, count data models, Argentina

JEL: R12; R30; C33

() Facultad de Ciencias Económicas y Sociales (Universidad Nacional de Mar del Plata) Funes 3250; 7600 - Mar del Plata (Argentina) Phone: + 54 223 474 9696 int. 338. Fax + 54 223 474 9696 () CREIP and Quantitative Urban and Regional Economics (QURE) Department of Economics (Universitat Rovira i Virgili) Av. Universitat, 1; 43204 – Reus (Catalonia, Spain) Phone: + 34 977 758 902. Fax + 34 977 759 810 (♦) Institut d’Economia de Barcelona (IEB) Av. Diagonal, 690; 08034 – Barcelona (Catalonia, Spain) Acknowledgements This paper was partially funded by SEJ2010-19088/ECON, SEJ2010-16934/ECON, the “Xarxa de Referència d’R+D+I en Economia i Polítiques Públiques” and the SGR programme (2009-SGR-322) of the Catalan Government. We would like to acknowledge research assistance by Magda Lleixà and Elisenda Jové. We also thank the Employment and Business Dynamics Observatory from Argentina for kindly providing the dataset.

1

1. Introduction

The new economic geography (Krugman, 1991; Venables, 2005) and the endogenous

growth theories (Aghion and Howitt, 1998) have both stressed the role of the spatial

distribution of economic activity in increasing development opportunities. In this

respect, the empirical evidence shows that firm dynamics may enhance regional job

growth (Ghani et al., 2011), increase the commercialization of innovations (Audretsch

et al., 2006), accelerate structural change (Gries and Naudé, 2010), and help discover

the competitive advantages of a nation (Hausmann and Rodrik, 2003). It is therefore

important to understand what determines the entry and exit of firms in developing

countries.

A number of previous studies have addressed these issues. However, most of them

have focused on the entry process. This is the case of Lay (2003) and Wang (2006) for

Taiwan, and Günalp and Cilasun (2006) and Ozturk and Kilic (2012) for Turkey, all of

whom analyse industry level data. Also, within the regional science literature we

should mention the studies by Naudé et al. (2008) for South Africa and Santarelli and

Tran (2012) for Vietnam. To our knowledge, the only studies on the aggregate

determinants of exit are the ones already mentioned by Lay (2003) for Taiwan and

Ozturk and Kilic (2012) for Turkey.1 This means that to date no empirical studies have

been made on the determinants of regional firm exit. This paper aims to fill this gap by

analysing the determinants of the (annual) number of exits in the Argentinean

provinces between 2003 and 2008.2

Of the developing countries, Argentina has a number of features that are worth noting.

First, it is a country with important regional differences in terms of wages, labour

skills, growth rates and natural resources. Ultimately, regional development levels

differ considerably across the country. Second, firms and people are highly

concentrated around the main urban areas, particularly the capital. Third, Argentina

2

covers a vast territory that is organised in large administrative units. Interestingly,

many other developing countries (e.g. South Africa, Brazil, Russia, Mexico and

Vietnam) to some extent share these features. This means that although it may not be

possible to generalise our results to all developing countries, they are likely to hold for

a number of them.

Our main finding is that the spatial distribution of exits exhibits a core-periphery

structure that is mostly driven by the effects of entrants, incumbents and the informal

economy. First, the so-called “revolving door effect” (Audretsch, 1995), by which past

entrants push firms out of the markets, is less intense in the central regions. Second,

peripheral regions with a strong industrial structure (proxied by the number of past

incumbents) and/or economic activity (proxied by the number of current incumbents)

have fewer exits than their counterparts in the central regions. Third, the informal

economy has a non-linear impact on exit. The effect is initially negative (i.e., the

larger the informal economy, the fewer exits there are). However, it becomes positive

when the size of the informal economy grows. So the informal economy increases the

number of exits, and more so in the peripheral regions.

The rest of the paper is organised as follows. Section 2 reviews the relevant literature.

It also discusses our model specification. Section 3 describes the data set. Section 4

discusses the econometric models and the main results. Section 5 concludes.

2. Literature review

2.1 Firm exit in developed countries

The industrial organization approach to the analysis of firm exit stems from the fact

that exits occur when the (expected) profit falls below a particular threshold

(Jovanovic, 1982; Ghemawat and Nalebuff, 1985; Klepper, 1996; Das and Das, 1996).

Thus, we expect that differences in exit rates among industries to be closely related to

3

differences in the proportion of firms with losses. Also, the higher the rate of industry

growth, the lower the number of exits will be, since more firms are expected to cover

their costs and realize profits. Lastly, the exit threshold depends on the extent of exit

barriers so exit rates are negatively related to the ratio of sunk to variable costs (Caves

and Porter, 1976, Mac Donald, 1986, Frank, 1988).

In the regional science literature, however, the emphasis lies on the characteristics of

the region where the firm is located (Baldwin et al., 2000).3 In particular, the

significant variations in regional exit patterns are mainly explained by differences in

regional labour markets, regional industrial composition, and the spatial concentration

of economic activities and individuals. As for the differences in the labour market, the

literature has concentrated on the effects of unemployment. On the one hand, an

increase in unemployment may have a negative impact on exit because self-employed

individuals face fewer job opportunities and are thus less prone to exit (Carree and

Thurik, 1996; Lin et al., 2001; Nyström, 2007a, 2007b; Carree et al., 2008; Santarelli

et al., 2009). On the other hand, unemployment is a proxy for the level of activity of

the economy and an increase may result in an increase in the number of exits

(Buzzelli, 2005; Brixy and Grotz, 2007; Fertala, 2008). As for the differences in

industrial composition, the lower the complexity and diversity of the local industrial

structure, the lower the ability to reallocate resources to new activities when a negative

shock occurs (Kosacoff and Ramos, 1999). Thus, exit is more likely in less diversified

environments. Lastly, since firms need to be close to other firms and workers to

benefit from agglomeration economies and market-oriented firms from physical

proximity to consumers, non-concentrated areas will tend to have more exits (Keeble

and Walker, 1994; Littunen et al., 1998). However, disagglomeration economies may

increase the production costs and lead to further exit of firms. This is because a higher

density pushes up input prices by increasing competition for the scare resources

(Agarwal and Gort, 1996; Huisman and van Wissen, 2004; Fritsch et al., 2006). Exits

4

may be higher in densely populated areas ―see e.g. Buss and Lin (1990), Forsyth

(2005) and Huiban (2011) for empirical evidence. There are several reasons for this.

First, competition in both goods and factor markets can be higher (Agarwal and Gort,

1996; Bresnahan and Reiss, 1991). Second, the chances of finding a job, finding an

entrepreneurial opportunity and/or selling the firms’ assets to another venture can be

higher (Huiban, 2011). Third, as discussed below, since large urban areas attract more

entry, the higher share of young firms may lead to higher exits.

At the aggregate level, exits have been shown to increase during downturns

(Audretsch and Mahmood, 1995; see, however, Boeri and Bellman, 1995). In

particular, the level of regional demand may be relevant for services and local-market

driven manufacturing activities. Also, we expect low real interest rates to discourage

firm exit (Kendall et al., 2010). These effects are particularly important for small

firms, which are generally more likely to exit due to cost disadvantages that make

them less able to compete efficiently and survive (Fotopoulos and Spence, 1998;

Esteve et al., 2004; Box, 2008; Carreira and Teixeira, 2011). Thus, the “liability of

smallness” means that exits should be higher in regions with a large proportion of

small firms. This is closely related to the “revolving door” phenomenon by which

many firms exit only a few years after creation (Audretsch, 1995). The displacement

effect of the new entrants has been empirically documented in developed countries

both at the industry and regional levels (Arauzo-Carod et al., 2007; Manjón-Antolín,

2010).

2.2 Firm exit in developing countries

We have just shown that there is an extensive empirical literature on regional firm

exit. In contrast, the evidence from developing countries is scarce. We should mention

the studies by Lay (2003) and Ozturk and Kilic (2012), who analyse the determinants

5

of sectorial exit in Taiwan and Turkey, respectively, and the studies by Frazer (2005),

Eslava et al., (2006), López (2006), Alvarez and Görg (2009) and Alvarez and

Vergara (2010; 2013), who seek to explain firm exit using size, age and productivity

as the main covariates. To our knowledge, this is the first study on the determinants of

regional firm exit.

In particular, we consider a set of determinants that are meant to replicate those

typically used in studies on developed economies (e.g., agglomeration economies).

However, we also acknowledge that there are factors that, while potentially important

in developing countries, are generally not considered by studies on developed

countries (e.g., the informal economy). This specification is rather ad-hoc, but it is

important to stress that there is no well-established theory that provides guidelines on

what the determinants of exit are in a developing country and on whether their

expected effects are (dis)similar to the expected effects in a developed country. With

this in mind, we argue that macroeconomic and financial factors can have a different

impact on exit in developing and developed countries, whereas structural factors can

have a different impact within the regions of a developing country (centre vs.

peripheral regions).

First, developing economies are generally characterised by macroeconomic instability

and intense cyclical variations (Stiglitz, 1998; Ocampo et al., 2009; Bértola and

Ocampo, 2012), so vulnerability to external (and internal) shocks is expected to be

higher. This means that after each crisis a considerable number of firms enter the

growing markets, many of which will exit in the following years (the greater the

decline, the more firms exit), thus producing a “revolving door” phenomenon that is

often more intense than in developed countries. In addition, the fact that economic

cycles are more pronounced in developing countries reinforces the anticyclicality of

exits. Because of the worse credit conditions in developing countries, high real interest

6

rates are also expected to discourage firm exit less than in developed countries

(Kendall et al., 2010). Lastly, developing countries show marked differences in critical

economic indicators among their regions, to the extent that some regions can have

levels of capitalization, technology, productivity, organization and human capital

requirements similar to their counterparts in advanced countries (Sunkel, 1978). A

direct implication of this “structural heterogeneity” (Cassiolato et al., 2009) is that

firm exit determinants may differ across the regions of a country.4

3. The data

3.1 Exit

Exit data used in this paper comes from the Employment and Business Dynamics

Observatory (EBDO) of the Ministry of Labour and Social Security of Argentina.

More specifically, the database includes information about the number of entries, exits

and incumbents based on all manufacturing (formal and private) firms with at least

one employee registered with the Social Security. This means that our data set does

not contain information on either public or informal employment. Moreover, the

EBDO handles changes in firm codes that do not reflect true market entries and exits.

In general, a firm is considered closed when it does not declare employees for a period

of twelve months. However, spurious exits caused by the displacement of a whole

firm’s workforce from firms that “exit” to become “new” firms have been identified

and excluded. Lastly, we restrict the analysis to firms that declare that most of their

workforce is located in the assigned jurisdiction. This means that branch offices or

subsidiaries located in other jurisdictions are excluded from our data set. All in all, this

is the most up-to-date, comprehensive, reasonably long-term and spatially

disaggregated data source currently available for firm demography studies in

Argentina.

7

Data is available for the 23 Argentinean provinces and the Capital Federal city. These

are our units of observation. However, the Buenos Aires Province is actually divided

into Gran Buenos Aires and the rest of the province. We also decided to drop the

province of Río Negro because of missing data in most of the explanatory variables

we considered. Therefore, although there are 25 jurisdictions in the database, we

ultimately provide results for only 24. Thus, our dependent variables are the number of

annual exits in each jurisdiction between 2003 and 2008. We start our analysis in 2003

to avoid the structural break caused by the economic and political crisis of the end of

2001 that led to the devaluation of the Argentinean peso in January 2002. Including

these years of turmoil would have completely distorted the results. We finish our

analysis in 2008 because this was the last year available in the EBDO dataset when

this investigation was initiated. Table 1 shows the evolution of entries, exits and

incumbents over the period of analysis.

[INSERT TABLE 1 HERE]

Exits followed an increasing path after the first two years of stability (2003-2004).

According to the MTEySS (2007), this was largely driven by new ventures after the

deep economic recession of 2000-2001 (deferred projects along with strictly new

ventures encouraged by better macroeconomic conditions). Thus, while entries in

2003-2005 doubled the entries in 2000-2002, exits increased at an average rate of 20%

after 2005. Additionally, the slowdown in the net entry in 2008 is explained by the

international financial crisis, the gradual appreciation of the real exchange rate and

some internal conflicts (Katz and Bernat, 2011).

Figure 1 shows that the spatial distribution of these exits is not homogeneous, since

most concentrate on the richest five regions: namely, the Capital Federal city and the

provinces of Gran Buenos Aires, the rest of Buenos Aires province, Santa Fe and

8

Córdoba. More precisely, these regions cover roughly 22% of the surface of the

country but concentrate about 80% of the workers, incumbent firms and exiting firms.

[INSERT FIGURE 1 HERE]

The existence of a different pattern of exit in the central and peripheral regions is

emphasized in Figure 2, where we plot the evolution of the number of exits in both

sets of provinces. Notice that both the levels of the variable and the slope of the curve

differ. Exits follow an increasing path in both sets of provinces, but at a higher rate in

the richest. The combined result is that the number of exits in Argentina practically

doubled during the period of analysis. Also, since the increase in the number of entries

was smaller, the population of firms shrank (see Table 1).

[INSERT FIGURE 2 HERE]

3.1 Explanatory variables

We used data from the EBDO and the National Household Survey (NHS) to construct

our vector of explanatory variables (the size of the provinces in km2 comes from the

Military Geographical Institute). The distinction is important because the information

contained in the EBDO database refers to the whole province, while the NHS is

performed by the National Institute of Statistics and Census (INDEC) on samples of

families in 31 urban areas (“aglomerados”). Nevertheless, we were obliged to use the

NHS data because no statistical source provides yearly information on demographic

and/or socioeconomic characteristics of the Argentinean provinces (population

censuses, for example, are performed every 10 years).5

In particular, we were able to construct variables related to the evolution of economic

activity, the labour market, the industrial structure, the existence of agglomeration

9

economies and the number of entries. As discussed in the section above, these factors

are widely used in studies on developed countries. We also included among the

covariates a measure of the informal economy, which is a structural singularity of the

developing countries (Schneider, 2005) and needs to be taken into account and the

square of this variable to account for non-linear effects. Lastly, we explored the

existence of core-periphery differences by including the products of a dummy that

identifies the richest provinces (the Capital Federal city, Gran Buenos Aires, the rest

of Buenos Aires province, Santa Fe and Córdoba) with all the regional determinants

previously mentioned. Year dummy variables were also included to control for

macroeconomic factors. These were preferred to macroeconomic variables such as e.g.

GDP growth because of the problems of measuring these aggregates. The GDP growth

in local currency is inaccurate because official inflation figures have not been reliable

since 2007 and the GDP growth in US dollars is similarly misleading because of the

severe devaluation of the Argentinean peso in 2002 (more than 200%) and the

consequent gradual appreciation. Notice also that we have not included measures of

credit access in our set of explanatory variables. Actually, we explored the use of the

number of loans granted i) to manufacturing, ii) per firm and iii) per employee.

However, these variables were statistically non-significant and results did not differ

substantially from the ones reported in Table 3. We consequently decided not to

include these variables in our final specifications.

Table 2 reports the definition, statistical sources and descriptive statistics of the

explanatory variables used in this study. We have included a column with the expected

sign of the associated coefficient. Next we briefly review the arguments and evidence

supporting these expected signs.

[INSERT TABLE 2 HERE]

10

Business cycle. We use the rate of variation of employment in all formal firms

(alternatively, the rate of variation of unemployment) to proxy for the evolution of

economic activity. The coefficient of this variable is expected to be negative (positive

for the rate of variation of unemployment), thus reflecting the procyclicality of exits.

Labour. We use wages and the unemployment rate to assess the impact of the labour

market on firm exit. Wages correspond to the average monthly wage of private

registered workers, in nominal terms because official inflation rates in Argentina have

not been reliable since 2007. We expect a positive sign for this variable. As for the

unemployment rate, we cannot say, a priori, what its impact on exit will be.

Industrial structure. The industrial structure of the province is approximated using the

Hirschman-Herfindahl Index, which measures lack of diversity. We expect this

variable to impact positively on exit, since firms located in less diversified

environments are more vulnerable to external shocks. We also control for the previous

industrial activities carried out in a province using the average number of incumbents

7, 6 and 5 years before (i.e. a 3-year centered moving average). We expect past

incumbents to have developed a favourable business environment and supporting

institutions that hamper exit. However, changes in the conditions that determine

profitability (the high macroeconomic volatility of developing countries affects the

exchange rate, credit conditions, tax policy, etc.) and the lack of continuity in

industrial policies can mitigate this effect.

Spatial concentration. We have included population density and its square, which have

been widely used as proxies for agglomeration and disagglomeration economies,

respectively. Both positive and negative signs are possible for the density coefficient,

11

while a positive sign is expected for its square. We have also included the number of

incumbent firms as an additional measure of the agglomeration of economic activity.

Entry rates. We use the (lagged) number of entries to account for the interdependence

between entries and exits. We expect this variable to show a positive coefficient.

Informal economy. We use the ratio of non-registered workers to total workers as a

proxy for the regional productive structure (e.g. the seasonality and/or low

productivity of certain activities may facilitate the growth of the informal sector)

and/or the lack of government controls over informal economy. The impact of this

variable on exit is ambiguous. A positive sign may arise if formal firms compete for

the same resources as informal firms and/or formal firms become informal when

facing difficulties. However, a negative sign is expected if formal firms buy inputs to

the informal sector, thus lowering costs and/or increasing flexibility.

4. Econometric modelling and estimation results

Given the definition of our dependent variable, we rely on panel count data models to

estimate the impact of exit determinants as have been done previously by some

scholars (Cameron and Trivedi, 1998; Ilmakunnas and Topi, 1999). Panel data models

were preferred to cross-section estimates on the grounds of two empirical tests. First,

likelihood ratio tests on the variance of the individual effects always yield statistically

significant results, thus rejecting the validity of pooled estimates (Cameron and

Trivedi, 2009). Second, we tested the assumption that observations are indeed

independent across the years studied by computing the covariance matrix of the year

vector of Pearson-residuals from the pooled Poisson regression model (see Hausman

et al., 1984 for details). We found large values in the off diagonal elements of the

matrix in all the specifications, which supports the independence assumption that

sustains panel data models.

12

It is also interesting to note that there are no zeros in the dependent variable. That is, in

each jurisdiction-year pair of our sample we have a strictly positive number of exits.

This is why we concentrate on the estimation of Poisson and Negative Binomial

models (Cameron and Trivedi, 1998). In particular, in Table 3 we report results from

the Poisson fixed effects model.6 Our choice is based on the results from a number of

tests (see the bottom rows of Table 3). First, the ratio of the Pearson goodness-of-fit

statistic to the degrees of freedom of a Poisson model with province dummy variables

is close to one. As Allison and Waterman (2002) argue, this indicates that there is no

overdispersion in the data. Second, the Durbin-Hu-Hausman test does not reject the

null hypothesis of no correlation between the covariates and the individual effect,

which means that the random effects model yields inconsistent estimates. Lastly,

Poisson fixed effects estimates provide the best fit according to the Akaike

Information Criterion (AIC).

[INSERT TABLE 3 HERE]

Let us first consider the results from the specification that contains variables which are

widely used in studies on developed countries. These are reported in the first column

in Table 3. What is most striking about these estimates is their lack of statistical

significance. In particular, only the level of wages and the measure of

disagglomeration economies show statistically significant coefficients. Also, these

findings hold when we include our proxy for the informal economy. These results are

reported in the second column of Table 3. The fit of the model is now better and our

measure of the informal economy and its square are both statistically significant.

However, the rest of the coefficients and their significance remain practically

unaltered.

13

We now go on to consider the results obtained when the cross-products of the regional

variables are included. Table 3 reports these results in the third column (without the

cross-products with the informal economy measures) and fourth column (with these

cross-products). The first thing to point out is that, unlike our previous specifications,

a number of variables are now statistically significant: namely, the rate of

unemployment, the number of lagged entries, the density measure, the industrial

tradition and the number of incumbents. In particular, the negative impact of the rate

of unemployment may be due to the lower cost of the workforce (not that so in the

central provinces) and/or reflect the small chances of finding a job by entrepreneurs

closing down their business (whereas the positive impact in the central provinces may

reflect that these chances are higher, as pointed out in footnote 4). Also, the negative

and positive coefficient of the density and its square is consistent with the existence of

(dis)agglomeration economies. On the other hand, wages are no longer significant.

The fit of the model, however, improves.

Moreover, the cross-products terms reveal that the spatial distribution of exits exhibits

a core-periphery structure whose main explanatory factors are the number of lagged

entrants, the number of past and current incumbents, and the size of the informal

economy (the unemployment rate and the HH index only matter when the informal

economy is not considered). In particular, rich provinces seem to be more able to

retain firms that are expelled from the markets by the new entrants. In other words, the

so-called “revolving door effect” is more intense in the other provinces. Also, there are

fewer exits in provinces that have a stronger industrial tradition (proxied by the

number of past incumbents) and more economic activity (proxied by the number of

current incumbents). Further, these effects are particularly strong in the less rich

provinces. Lastly, the existence of a small informal economy in the province prevents

exit. This may be related to the lower costs and/or higher flexibility that are inherent to

the informal hiring. However, the informal economy hastens exit when it grows

14

beyond a certain level, and, therefore, starts competing for resources also exploited by

formal firms.

We conclude by noting that our results are robust to alternative specifications of the

model. In particular, we dropped the number of two-year lagged entries (i.e., we

estimated the model including only the entries lagged one year), replaced the rate of

variation of employment in all formal firms by the rate of variation in unemployment ,

replaced density and its square by the ratio between the population in the main urban

areas of the province (“aglomerados”) and the total population of the province, and

replaced the rate of variation of employment in all formal firms by the rate of variation

in unemployment and density and its square by the ratio between the population in the

main urban areas of the province (“aglomerados”) and the total population of the

province. Estimates from these alternative specifications (available upon request)

show that although some of the coefficients vary in value and/or statistical

significance with respect to those reported in Table 3, most of the conclusions still

hold.

5. Conclusions

There is an extensive empirical literature on firm exit. However, little is known about

the determinants of firm exit in developing countries. This paper aims to fill this gap

in the literature by analysing the impact of regional factors on the yearly number of

exits in the Argentinean provinces using panel count data models. We find that while

past entrants increase current exits mostly in the peripheral regions, current and past

incumbents cause an analogous displacement effect but mostly in the central regions.

We also find that there is a U-shaped relationship between exits and the informal

economy, particularly in the peripheral regions.

15

In general, these findings can be useful for policy makers seeking to prevent firms

from exiting in certain areas. But they can also be helpful in the implementation and

evaluation of related policies. To illustrate, entry promoting policies can be used as an

instrument to boost economic activity in the more depressed areas. However, our

results indicate that such policies may ultimately cause more exits. This negative side

effect should thus be taken into account when assessing the welfare implications of

these policies. Also, it has been suggested that the informal economy hampers the

economic development of the (lagged) regions. However, our results indicate that,

these concerns are justified with regard to exits only when the size of the informal

economy is substantial. Moderate levels of informality, on the other hand, should not

be a major concern.

As for the future extensions of this work, we can name at least two. First, we will

explore the use a more disaggregated unit of observation. Given the lack of reliable

data on smaller geographical units (municipalities, counties and/or metropolitan

areas), using a sectorial breakdown will not only allow us to incorporate industry-

specific variables but to reduce the degree of heterogeneity in the regional units. We

will also explore the differences between exit rates of firms of different sizes. This can

be seen as a way to incorporate one of the main firm-level factors that determines exit.

16

References

Agarwal, R., & Gort, M. (1996). The Evolution of Markets and Entry, Exit and

Survival of Firm., The Review of Economics and Statistics, 78, 489-498.

Aghion, P., & Howitt, P. (1998). Endogenous Growth Theory. MIT Press.

Allison, P.D., & Waterman, R.P. (2002). Fixed-Effects Negative Binomial Regression

Models. Sociological Methodology, 3, 247-265.

Álvarez, R., & Görg, H. (2009). Multinationals and plant exit: Evidence from Chile.

International Review of Economics & Finance, 18 (1), 45-51.

Álvarez, R., & Vergara, S. (2013). Trade exposure, survival and growth of small and

medium-size firms. International Review of Economics & Finance, 25, 185-201.

Álvarez, R., & Vergara, S. (2010). Exit in Developing Countries: Economic Reforms

and Plant Heterogeneity. Economic Development and Cultural Change, 58 (3), 537-

561.

Arauzo-Carod, J. M., Manjón-Antolín, M., Martín, M., & Segarra, A. (2007).

Regional and sector-specific determinants of industry dynamics and the displacement–

replacement effects. Empirica, 34 (2), 89-115.

Audretsch, D. B. (1995): Innovation and industry evolution. The MIT Press.

Audretsch, D. B., Keilbach, M. C., & Lehmann, E. E. (2006). Entrepreneurship and

economic growth. USA: Oxford University Press.

17

Audretsch, D. B., & Mahmood, T. (1995). New Firms Survival: New Results Using a

Hazard Function. The Review of Economics and Statistics, 77 (1), 97–103.

Baldwin, J. R. B., Dupuy, L., & Richard Gellatly, G. (2000). Failure Rates for New

Canadian Firms: New Perspectives on Entry and Exit.  Ottawa, Canada: Statistics

Canada.

Bartelsman, E., Haltiwanger, J., & Scarpetta, S. (2004). Microeconomic Evidence of

Creative Destruction in Industrial and Developing Countries. Institute for the Study of

Labor, IZA, Discussion Paper nº 1374. The World Bank, Policy Research Working

Paper nº 3464.

Bértola, L., & Ocampo, J.A. (2012). The Economic Development of Latin America

since Independence. Oxford University Press.

Boeri, T., & Bellman, L. (1995). Post-Entry Behaviour and the Cycle: Evidence from

Germany. International Journal of Industrial Organization, 13 (4), 483-500.

Box, M. (2008). The death of firms: exploring the effects of environment and birth

cohort on firm survival in Sweden. Small Business Economics, 31, 379-393.

Bresnahan, T. F., & Reiss, P. C. (1991). Entry and competition in concentrated

markets. Journal of Political Economy, 99 (5), 977-1009.

Brixy, U., & Grotz, R. (2007). Regional patterns and determinants of birth and

survival of new firms in Western Germany. Entrepreneurship & Regional

Development, 19 (4), 293-312.

18

Buss, T. F., & Lin, X. (1990). Business survival in rural America: A three-state study.

Growth and Change, 21, 1–8.

Buzzelli, M. (2005). What explains firm transience in house-building? A regional

analysis of Ontario, Canada, 1991 and 1996. Regional Studies, 39 (6), 699–712.

Calá, C.D., Manjón-Antolín, M., & Arauzo-Carod, J.M. (2012). Regional

Determinants of Firm Entry in a Developing Country. Working paper nº 18,

Departament d’Economia & CREIP (Universitat Rovira i Virgili).

Calá, C.D., & Rotondo, J.S. (2012). Dinámica empresarial en la industria argentina.

Un análisis provincial para el período 2003-2008. Revista Realidad Económica, 267,

138-170.

Cameron, A.C., & Trivedi, P.K (2009). Microeconometrics Using Stata, Stata Press.

Cameron, A.C., & Trivedi, P.K (1998). Regression Analysis of Count Data,

Cambridge: Cambridge University Press.

Carree, M., Santarelli, S., & Verheul, I. (2008). Firm entry and exit in Italian

provinces and the relation-ship with unemployment. International Entrepreneurship

Management Journal, 4, 171–186.

Carree M., & Thurik R (1996). Entry and exit in retailing: incentives, barriers,

displacement and replacement. Review of Industrial Organization, 11 (2), 155-172.

19

Carreira, C., & Teixeira, P. (2011). The shadow of death: analysing the pre-exit

productivity of Portuguese manufacturing firms. Small Business Economics, 36, 337–

351.

Cassiolato, J.E., Pagola, C.B., & Lastres, H.M. (2009). Technical Change and

Structural Inequalities: Converging Approaches to Problems of Underdevelopment. In

W.J. Drechsler, R. Kattel, & E.S., Reinert (Eds.), Techno-Economic Paradigms:

Essays in Honour of Carlota Perez. Anthem Press.

Caves, R., & Porter, M. (1976). Barriers to exit. In R. Masson and D. Qualls (Eds.),

Essays on industrial organization in honour of Joe Bain, Cambridge, MA: Ballinger.

Das, S., & Das, S. (1996). Dynamics of entry and exit of firms in the presence of entry

adjustment costs. International Journal of Industrial Organisation, 15, 217-241.

ECLAC (Economic Commission for Latin America and the Caribbean) (2002). The

integration of Latin America and the Caribbean in global trade and production circuits.

In Globalization and Development, Ch. 6, ECLAC.

Eslava, M., Haltiwanger, J.C., Kugler, A.D., & Kugler, M. (2006). Plant turnover and

structural reforms in Colombia. IMF Staff Papers, 53, 58–75.

Esteve, S., Sanchís, A., & Sanchís, J. (2004). The Determinants of Survival of Spanish

Manufacturing Firms. Review of Industrial Organization, 25, 251-273.

Fertala, N. (2008). The shadow of death: do regional differences matter for firm

survival across native and immigrant entrepreneurs?. Empirica, 35, 59–80.

20

Fotopoulos, G., & Spence, N. (1998). Accounting for Net Entry into Greek

Manufacturing by Establishments of Varying Size. Small Business Economics, 11,

125–144.

Frank, M. (1988). An Intertemporal Model of Industrial Exit. The Quarterly Journal

of Economics, 103 (2), 333-344.

Frazer, G. (2005). Which firms die? A look at manufacturing firm exit in Ghana.

Economic Development and cultural change, 53 (3), 585-617.

Forsyth, G. D. (2005). A note on small business survival in rural areas: The case of

Washington State. Growth and change, 36, 428–440.

Fritsch, M., Brixy, U., & Falck, O. (2006). The effect of industry, region, and time on

new business survival. Review of Industrial Organization, 28, 285-306.

Ghani, E., Kerr, W. R., & O’Connell, S. (2011). Promoting Entrepreneurship,

Growth, and Job Creation. Edited by Ghani, E. Ch. 7, 166-199.

Ghemawat, P., & Nalebuff, B. (1985). Exit. RAND Journal of Economics, 16, 184-

194.

Günalp, B., & Cilasun, S.M. (2006). Determinants of Entry in Turkish Manufacturing

Industries. Small Business Economics, 27, 275–287.

Gries, T., & Naudé, W. (2010). Entrepreneurship and structural economic

transformation. Small Business Economics, 34(1), 13-29.

21

Hausman, J., Hall, BH., & Griliches, Z. (1984), Econometric Models for Count Data

with an Application to the Patents-R & D Relationship. Econometrica, 52 (4), 909-

938.

Hausmann, R., & Rodrik, D. (2003). Economic development as self-discovery.

Journal of Development Economics, 72 (2), 603–633.

Huiban, J. P. (2011). The spatial demography of new plants: urban creation and rural

survival. Small Business Economics, 37 (1), 73-86.

Huisman, C., & van Wissen, L. (2004). Localization effects of firm startups and

closures in the Netherlands. Annals of Regional Science, 38, 291-310.

Ilmakunnas, P., & Topi, J. (1999). Microeconomic and Macroeconomic Influences on

Entry and Exit of Firms. Review of Industrial Organization, 15, 283–301.

Jovanovic, B. (1982). Selection and the evolution of industry. Econometrica, 50 (3),

649–70.

Katz, J., & Bernat, G. (2011). Exit-entry, productivity growth and structural change in

response to changes in macroeconomic policy: evidence from Argentina. Innovation

and Development, 1 (2), 227-244.

Keeble, D., & Walker, S. (1994). New Firms, Small Firms and Dead Firms: Spatial

Patterns and Determinants in the United Kingdom. Regional Studies, 28 (4), 411-427.

Kendall, J., Mylenko, N., & Ponce, A. (2010). Measuring Financial Access around the

World. Policy Research Working Paper 5253, The World Bank, March.

22

Klepper, S. (1996). Entry, exit, growth and innovation over the product cycle.

American Economic Review, 86 (3), 562–583.

Kosacoff, B., & Ramos, A. (1999). El debate sobre política industrial. Revista de la

CEPAL, 68, 35-60.

Krugman, P. (1991). Increasing returns and economic geography. Journal of Political

Economy, 99, 483-499.

Lay, T.J. (2003). The Determinants of and Interaction between Entry and Exit in

Taiwan´s manufacturing. Small Business Economics, 20, 319-334.

Lin, Z., Picot, G., & Compton, J. (2001). The entry and exit dynamics of self-

employment in Canada. Small Business Economics, 15, 105–125.

Littunen, H., Storhammar, E., & Nenonen, T. (1998). The survival of firms over the

critical first 3 years and the local environment. Entrepreneurship & Regional

Development, 10 (3), 189-202.

López, R. A. (2006). Imports of intermediate inputs and plant survival. Economics

Letters, 92 (1), 58-62.

Love, J.H. (1996). Entry and exit: a county-level analysis. Applied Economics 28, 441-

451.

MacDonald, J.M. (1986). Entry and Exit on the Competitive Fringe. Southern

Economic Journal, 52, 640-652.

23

Manjón-Antolín, M. C. (2010). Firm size and short-term dynamics in aggregate entry

and exit. International Journal of Industrial Organization, 28 (5), 464-476.

MTEySS (2007). La creación de empresas durante la etapa post convertibilidad (2003-

2005): Impacto sobre el empleo asalariado registrado. In Salarios, empresas y empleo

2003-2006, Series Trabajo, ocupación y empleo, nº 5. National Ministry of Labour,

Employment and Social Security. January.

Naudé, W., Gries, T., Wood, E., & Meintjies, A. (2008). Regional determinants of

entrepreneurial start-ups in a developing country. Entrepreneurship & Regional

Development: An International Journal, 20 (2), 111-124.

Nyström, K. (2007a). An industry disaggregated analysis of the determinants of

regional entry and exit. Annals of Regional Science, 41, 877–896.

Nyström, K. (2007b). Patterns and determinants of entry and exit in industrial sectors

in Sweden. Journal of International Entrepreneurship, 5, 85–110.

Ocampo, J. A., Rada, C., & Taylor, L. (2009). Growth and Policy in Developing

Countries: A Structuralist Approach. New York: Oxford University Press.

Ozturk, S., & Kilic, C. (2012). Patterns and determinants of entry and exit in Turkish

manufacturing industries. International Journal of Arts and Commerce, 1 (5), 107-

118.

Santarelli, E.; Carree, M., & Verheul, I. (2009). Unemployment and Firm Entry and

Exit: An Update on a Controversial Relationship. Regional Studies, 43 (8), 1061-1073.

24

Santarelli, E., & Tran, H.T. (2012). Growth of Incumbent Firms and Entrepreneurship

in Vietnam. Growth and Change, 43 (4), 638–666.

Schneider, F. (2005). Shadow economies around the world: what do we really know?.

European Journal of Political Economy, 21, 598–642.

Stiglitz, J. (1998). Towards a New Paradigm for Development: Strategies, Policies,

and Processes. Given as the 1998 Prebisch Lecture at UNCTAD, Geneva, October 19.

Sunkel O. (1978). La dependencia y la heterogeneidad estructural. Trimestre

Económico, 45 (1), 3-20 México, D.F.

Venables, A. J. (2005). Spatial disparities in developing countries: cities, regions, and

international trade. Journal of Economic Geography, 5 (1), 3-21.

Wang, S. (2006). Determinants of New Firm Formation in Taiwan. Small Business

Economics, 27, 313–321.

25

Table 1. Number of entries, exits and incumbents in Argentina (2003 – 2008)

Year Entry Exit Incumbents

2003 4,986 2,330 42,754

2004 5,994 2,326 45,234

2005 5,486 2,929 48,317

2006 6,264 3,623 49,987

2007 5,886 4,358 51,796

2008 5,389 5,103 52,417

Source: authors from data in EBDO

26

Figure 1. Number of exits by province (2003-2008 mean)

Source: authors from EBDO data. “GBA” stands for Gran Buenos Aires and “Bs As Rest” for the rest of the Buenos Aires province.

Misiones

Entre Ríos

Bs As Rest La Pampa

Río Negro

Chubut

Santa Cruz

Tierra del Fuego

Neuquén

Mendoza

San Luis

Córdoba

Santa Fe

Corrientes

Formosa

Chaco Santiago

del Estero

San Juan

Catamarca

La Rioja

Jujuy

Salta

Tucumán

GBA

Capital Federal

27

0

500

1000

1500

2000

2500

3000

3500

4000

4500

2003 2004 2005 2006 2007 2008

Number of firm

s

Central  regions Peripheral  regions

Figure 2. Number of exits in central and peripheral regions (2003-2008) Source: authors from EBDO data. Central regions include: Capital Federal city, Gran Buenos Aires, the rest of the Buenos Aires province, Córdoba and Santa Fe. Peripheral regions include: Catamarca, Chaco, Chubut, Corrientes, Entre Ríos, Formosa, Jujuy, La Pampa, La Rioja, Mendoza, Misiones, Neuquén, Río Negro, Salta, San Juan San Luis, Santa Cruz, Santiago del Estero, Tierra del Fuego and Tucumán.

28

Table 2. Explanatory variables: definition, sources, expected signs and descriptive statistics Variable Definition Source Expected sign Mean St. Dev. Min. Max Employment variation Rate of variation in employment in all formal firms - 9.22 5.20 -6.97 22.75

Wages Average monthly wage of private registered workers in manufacturing

Own calculations from EBDO +

1,891.40 864.87 676.17 5.414.11

Unemployment rate Unemployment rate Own calculations

from National Household Survey* +, -

8.19 3.81 1.01 18.20Entryt-2 Number of entries in the previous year (2 lags) + 190.85 342.43 3.00 1,609.00Entryt-1 Number of entries in the previous year (1 lag) + 212.04 368.99 3.00 1,609.00HH Index Hirschman-Herfindahl Index + 24.36 12.00 8.06 62.90Industrial tradition Incumbent firms 7 years ago (3-years moving average)

Own calculations from EBDO

- 1,916.31 3,396.97 91.00 14,550.00Density ln(Population/Area) (in thousands) +,- 2.63 2.06 -0.18 9.53

Density2 Density2

Own calculations from Military Geographical Institute and

INDEC + 11.14 20.38 0.01 90.78Incumbents Number of incumbent firms in the current year Own calculations from EBDO + 1,999.11 3,472.29 88.00 15,107.00

Informal Economy Non registered workers over registered workers Own calculations

from National Household Survey* +, -

0.81 0.31 0.16 1.51 * Data refer to 3rd quarter of every year, except for 2007 (4th quarter).

Source: authors

29

Table 3. Determinants of firm exit

[1] [2] [3] [4] -0.0073 -0.0053 -0.0081 -0.0073

Employment variation (0.0048) (0.0049) (0.0057) (0.0057) 0.0002*** 0.0002* 0.0001 0.0001

Wages (0.0001) (0.0001) (0.0001) (0.0001) -0.0080 -0.0058 -0.0310** -0.0279*

Unemployment rate (0.0081) (0.0082) (0.0150) (0.0152) 0.0001 0.0001 0.0028** 0.0028**

Entry t-2 (0.0001) (0.0001) (0.0012) (0.0012) 0.0001 0.0001 0.0039*** 0.0039***

Entry t-1 (0.0002) (0.0002) (0.0013) (0.0013) -0.0002 0.0049 -0.0084 -0.0047

HH Index (0.0092) (0.0094) (0.0154) (0.0155) -0.0000 -0.0000 -0.0015* -0.0018**

Industrial Tradition (0.0001) (0.0001) (0.0008) (0.0008) -3.9590 -3.5190 -5.7524* -7.2910**

Density (2.4706) (2.4777) (3.4532) (3.5629) 0.5278** 0.6170*** 1.4321*** 1.6811***

Density2 (0.2106) (0.2131) (0.4572) (0.4758) -0.0002 -0.0002 -0.0043*** -0.0046***

Incumbents (0.0002) (0.0002) (0.0012) (0.0012)

-1.1788** -1.0399** -1.6164** Informal Economy

(0.4645) (0.5104) (0.7036) 0.6774*** 0.8469*** 1.1888***

Informal Economy2 (0.2381) (0.2545) (0.3304) 0.0062 0.0030

Employment variation × Centre (0.0122) (0.0124) 0.0002 0.0001

Wages × Centre (0.0002) (0.0002) 0.0315* 0.0230

Unemployment rate × Centre (0.0189) (0.0194) -0.0026** -0.0035***

Entry t-2 × Centre (0.0012) (0.0013) -0.0035*** -0.0026**

Entry t-1 × Centre (0.0013) (0.0012) -0.0948** -0.0283

HH Index × Centre (0.0456) (0.0613) 0.0015* 0.0017**

Industrial Tradition × Centre (0.0008) (0.0008) -4.8424 0.9861

Density × Centre (10.1453) (10.5126) -0.1779 -0.8311

Density2 × Centre (0.7013) (0.7696) 0.0039*** 0.0041***

Incumbents × Centre (0.0012) (0.0012) 1.5760

Informal Economy × Centre (1.0540) -1.1534*

Informal Economy2 × Centre (0.5886)

AIC 773.06 768.69 767.27 766.95 LR Test of Joint Significance 1797.82*** 1805.78*** 1827.64*** 1829.63*** Hausman 16.24* 36.99*** 51.22*** 42.01*** Pearson’GoF Test 108.19 99.87 78.47 74.35

Note: Poisson Fixed Effects estimates are reported. Standard errors in brackets. Asterisks indicate the statistical significance of the coefficient: *** p-value < 0.01, ** p-value < 0.05, * p-value < 0.1. Year dummy variables are included in all the specifications.

30

1 A number of studies also use firm level data: e.g. Frazer (2005) for Ghana, Eslava et al. (2006) for Colombia,

and López (2006), Alvarez and Görg (2009) and Alvarez and Vergara (2010; 2013) for Chile.

2 Previous studies of firm exit in Argentina are merely descriptive (Bartelsman et al., 2004; MTEySS, 2007;

Katz and Bernat, 2011; Calá and Rotondo, 2012). Of these, Calá and Rotondo (2012) is the only one that adopts

a regional approach.

3 We use the terms “region” and “area” to refer to any geographical unit within a country. They are therefore not

necessarily linked to administrative units (e.g., regions, provinces, etc.). However, most of the studies considered

in this section use NUTS-II levels (i.e., regional level) and only a few smaller units (e.g., counties, as in the case

of Love, 1996).

4 This “structural heterogeneity” has accentuated in recent years: while there are now many more “world-class”

firms in developing countries, there is also a growing proportion of employment concentrated in low-

productivity informal-sector activities (ECLAC, 2002).

5 See e.g. Calá et al. (2012) for details on the construction of this data set.

6 The coefficient estimates in Table 3 can be interpreted as semi-elasticities. We do not report marginal effects

because of the difficulties in integrating out the unobserved heterogeneity in non-linear models (Cameron and

Trivedi, 2009).